Midterm 2 Study Guide POS3713 Does our independent variable really cause the change we see in dependent variable Can you demonstrate X causes Y Internal Validity External Validity Represents the degree to which we can be confident that the results of our analysis apply not only to the participants in the study but also to the population more broadly construed or Do your results from your study apply to other areas Good Research Designs Account for The Unit of Interest What is it I am studying Temporal spatial variation of interest Cross Sectional Time Series May compare values of Y across units May compare values of Y across time Threats to Causality History events that could effect causality Maturation is the unit of interest changing by itself Testing they know it s a test Selection selecting people Regression to the mean the idea that there is a stable level you should be at Good Research design ensures internal validity Random assignment ensures internal validity Random controls for thins I can measure and things I can t measure Experimental Designs Field Experiments Natural Experiments Statistical Control Reliability Test Retest Method Alternative Form Split Halves Randomly assign individuals into groups but perform the manipulation in the real word best on external validity An outside event separates people into control and treatment groups not really an experiment Find all the measures for alternative variables that wouldn t be X and include them in your model Apply the same measurement to observations at different periods of time problem with method difference in time Using two different measures of the same concept at two different times Same problem with Test Retest difference in time Collect two different measures of the same concept at the same time This way you know the concept being measured has not changed over time Bias not random Bias systematic over or underreporting of values Reliability is so important we d rather have systematically biased measures them unreliable measures for hypothesis testing Validity Face Validity Content Validity elements Construct Validity Ordinal categories median Interval Does the measure accurately represent our concept Example What about asking how prejudiced someone about race Your really measuring the willingness to reveal prejudice and not prejudice Would reasonable people agree What are the essential elements to that concept and do we have those Is our concept associated with other measures that it should be related to Levels of Measurement Nominal The mathematical quality of the scores of a variable Scores are labels categories only they are not numbers They can not be ranked or operated on by any mathematical function Categories must be mutually exclusive and exhaustive categories no rankings can only use the mode Scores have some numerical quality and can be ranked though they are still The distance between those ranked is undefined Categories with ranking and gaps between ranking are meaningless mode Even intervals between ranks Zero point is arbitrary year temp Real numbers gaps are meaningful zero does not mean anything uses mean median mode Ratio Even intervals between ranks with a true zero point Zero means something s uses mean median mode Descriptive Statistics populations and samples Any well defined set of units of analysis Samples are drawn from a theoretically constituted population Sample parameters are estimates of population parameters The large the sample the smaller the sampling error this means our estimates of population parameters are more precise When dealing with a dataset your first step should be to summarize your data these summaries are called descriptive statistics The most suitable descriptive is mode Nominal Ordinal Nominal variables can be described on their frequency How many cases fall into a particular category With ordinal variables we can describe the data in other ways Mode most frequent value of a variable in a dataset Median when data arranged from lowest to highest middle value With interval ratio level variables we can describe the moments of the variable The moments describe the central tendency of a variable and the distribution of values around it Zero Sum Property Adds up all the deviations from the mean and the answer is always zero Least Square Property Take all the values subtract from the mean Expected Value If we did this an infinite number of times this is what we expect guess the mean always Distributions Normal Bell Curve Negative Skew Positive Skew Few scores at the lower end of the scale tail to the left Few scores at the upper end of the scale tail to the right Sample Variance Sample Standard Deviation Standard Error of the Mean Population Variance Population Standard Deviation
View Full Document
Unlocking...